Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Master guide for using Agent Code effectively. Includes configuration templates, prompting strategies "Thinking" keywords, debugging techniques, and best practices for interacting with the agent.
.claude/skills/dokhacgiakhoa-agent-code-guide/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -78% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -60% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -73% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -87% | 0% |
To provide a comprehensive reference for configuring and using Agent Code (the agentic coding tool) to its full potential. This skill synthesizes best practices, configuration templates, and advanced usage patterns.
CLAUDE.md)When starting a new project, create a CLAUDE.md file in the root directory to guide the agent.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 15,066 | 10,979 | -27% | 1 | 1 | 0% | 2,424 | 1,801 | -26% | 0 | 0 | — |
case-02 | fail→pass | 24,875 | 17,081 | -31% | 1 | 1 | 0% | 3,811 | 3,130 | -18% | 0 | 0 | — |
case-03 | fail→fail | 16,771 | 16,132 | -4% | 1 | 1 | 0% | 2,841 | 2,628 | -7% | 0 | 0 | — |
case-04 | pass→pass | 7,741 | 3,297 | -57% | 1 | 1 | 0% | 1,265 | 640 | -49% | 0 | 0 | — |
case-05 | fail→pass | 15,842 | 2,390 | -85% | 1 | 1 | 0% | 2,434 | 526 | -78% | 0 | 0 | — |
case-06 | fail→pass | 6,193 | 2,505 | -60% | 1 | 1 | 0% | 1,113 | 440 | -60% | 0 | 0 | — |
case-07 | fail→pass | 10,571 | 1,942 | -82% | 1 | 1 | 0% | 1,800 | 486 | -73% | 0 | 0 | — |
case-08 | fail→pass | 17,575 | 1,671 | -90% | 1 | 1 | 0% | 3,126 | 416 | -87% | 0 | 0 | — |
case-09 | pass→pass | 11,656 | 1,835 | -84% | 1 | 1 | 0% | 1,584 | 521 | -67% | 0 | 0 | — |
case-10 | fail→pass | 12,725 | 1,641 | -87% | 1 | 1 | 0% | 2,073 | 434 | -79% | 0 | 0 | — |
case-11 | fail→pass | 15,742 | 1,824 | -88% | 1 | 1 | 0% | 2,153 | 461 | -79% | 0 | 0 | — |
case-12 | fail→pass | 9,468 | 1,624 | -83% | 1 | 1 | 0% | 1,685 | 460 | -73% | 0 | 0 | — |
case-13 | pass→pass | 8,351 | 3,680 | -56% | 1 | 1 | 0% | 1,647 | 594 | -64% | 0 | 0 | — |
case-14 | pass→pass | 9,464 | 5,884 | -38% | 1 | 1 | 0% | 1,673 | 1,248 | -25% | 0 | 0 | — |
case-15 | pass→pass | 10,883 | 11,521 | +6% | 1 | 1 | 0% | 1,941 | 1,828 | -6% | 0 | 0 | — |
case-16 | fail→pass | 8,829 | 2,125 | -76% | 1 | 1 | 0% | 1,523 | 505 | -67% | 0 | 0 | — |
case-17 | pass→pass | 12,539 | 12,386 | -1% | 1 | 1 | 0% | 2,231 | 2,339 | +5% | 0 | 0 | — |
case-18 | fail→pass | 11,007 | 1,999 | -82% | 1 | 1 | 0% | 1,588 | 471 | -70% | 0 | 0 | — |
case-19 | pass→pass | 7,697 | 3,103 | -60% | 1 | 1 | 0% | 1,457 | 642 | -56% | 0 | 0 | — |
case-20 | pass→pass | 4,922 | 3,706 | -25% | 1 | 1 | 0% | 794 | 763 | -4% | 0 | 0 | — |
case-21 | pass→pass | 5,523 | 3,377 | -39% | 1 | 1 | 0% | 898 | 648 | -28% | 0 | 0 | — |
case-22 | pass→pass | 5,058 | 3,175 | -37% | 1 | 1 | 0% | 724 | 681 | -6% | 0 | 0 | — |
case-23 | pass→pass | 11,800 | 8,775 | -26% | 1 | 1 | 0% | 1,874 | 1,747 | -7% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +43 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.